Blondell J M
U.S. Environmental Protection Agency.
Cancer Detect Prev. 1988;11(3-6):209-23.
Urban-rural factors associated with cancer mortality were studied in Kentucky to develop likely hypotheses on cancer prevention. Factor analysis was performed on county data for 25 variables which selected 17 variables for use in a multiple regression analysis. Age-adjusted cancer mortality rates used as outcome variables in this analysis included total and colorectal cancer (male, female, and both sexes combined) and female breast cancer. Each of these cancer sites was regressed against the 17 independent variables. Best predictive models for cancer outcome were selected on the basis of maximum total variance explained. Subsistence farming, where most of the farm produce is kept for home use rather than sold, was the single best predictor of low overall cancer rates for males, females, and both sexes combined in rural and Appalachian Kentucky. The subsistence farming diet is one based largely on milk and whole grains, rich in the following suspected anticarcinogens: selenium, magnesium, calcium, protease inhibitors, fiber, phenolics, and allelopaths. The diet lacks variety in fruits and vegetables and is low in vitamins A and C.
为了提出可能的癌症预防假说，对肯塔基州城乡地区与癌症死亡率相关的因素进行了研究。对25个变量的县级数据进行了因子分析，从中选取了17个变量用于多元回归分析。本分析中用作结果变量的年龄调整癌症死亡率包括总癌症死亡率和结直肠癌死亡率（男性、女性以及男女合计）以及女性乳腺癌死亡率。将这些癌症发病部位的每一个与17个自变量进行回归分析。根据最大可解释总方差选择癌症结果的最佳预测模型。在肯塔基州农村和阿巴拉契亚地区，自给农业（即大部分农产品留作家庭自用而非出售）是男性、女性以及男女合计总体癌症低发病率的唯一最佳预测因素。自给农业饮食主要以牛奶和全谷物为基础，富含以下可疑抗癌物质：硒、镁、钙、蛋白酶抑制剂、纤维、酚类化合物和化感物质。这种饮食缺乏水果和蔬菜的多样性，维生素A和C含量较低。
